Training Module 2:

Snow Detection in
Weather Satellite Imagery

Adapted from An Introduction to Satellite Image Interpretation, Eric D. Conway and the Maryland Space Grant Consortium, ©1997, Johns Hopkins University Press, Baltimore, 255 pp with Interactive CD-ROM.

For more information about this book and how to order copies go to the JHU Press On-line Catalog

With improvements in resolution and a high frequency of coverage, NOAA polar orbiters and GOES satellites provide us with excellent tools to observe snow cover on land surfaces and investigate the effects of this snow cover on climate, soil moisture, flood frequency, and more. Detecting snow in satellite imagery can be quite challenging, however, as snow cover and cloud cover often exhibit similar characteristics in visible and infrared imagery.



In this section you will learn about some of the patterns to look for to separate snow cover from cloud cover in satellite imagery. You will then be provided with several image that contain both cloud and snow cover. Carefully examine each image and for each, determine where the snow cover is and where the cloud cover is located.

Full-sized imagery for this activity

Midwestern snow cover (near Kansas, Arkansas, Nebraska) [snow1.gif]
Snow and cloud cover in Oklahoma/Texas panhandle region [snow2.gif]
Snow and cloud cover over Mississippi River Valley [snow3.gif]
False color-enhanced image of snow cover following 1996 east coast blizzard [snow4.jpg]
Snow and cloud cover over Appalachians in Pennsylvania and New York [snow5.gif]
Snow cover following the 1979 blizzard [bliz_79.gif]


Using data from meteorological satellites, you can create snow cover maps for North America. As you collect satellite images, draw in the extent of snow cover on a blank base map of the North American continent. Once you identify the snow in the image, you have located the snow line. The area of the snow cover (in square kilometers) can be estimated with this chart. These data can then be stored and compared with data for different months, years, and locations. Seasonal changes in snow cover can be determined, and yearly snow cover data can be examined for evidence of global climate change.

When producing such a map using satellite imagery, it is important to distinguish snow cover from other features. Snow on the surface appears bright white to light gray in VIS imagery, depending on the nature of the terrain, the vegetation, the age of the snow, and the illumination angle. Since fresh snow can appear very white and resemble cloud cover, it is important to be able to distinguish between the two. It is helpful to examine a series of image shown in motion; the clouds usually move, while snow does not.

Knowledge of the terrain can also help identify snow. Snow cover ends at unfrozen rivers or lakes, while clouds would likely cover these features. Snow on broad, flat, treeless areas (such as tundra) appears uniformly white, resembling clouds. Without clues such as rivers or lakes, it may not be possible to distinguish between snow and cloud cover. In areas covered by extensive forests, a snow-covered surface appears spotted, with lighter shades in treeless areas, and darker shades in wooded areas where the trees are thick and obscure the snow. Mountain snow can often be easily recognized by the dendritic pattern that forms due to lighter, snow-covered ridges surrounded by darker, tree-filled valleys and/or hillsides.

Shadows in VIS imagery can also aid in separating cloud cover from snow. Clouds can be seen casting shadows on the snow-covered surface in satellite imagery. IR imagery can also be enhanced to highlight the temperatures associated with snow and ice on the Earth's surface in order to locate these conditions in nighttime IR satellite imagery.

This visible image of the eastern United States was taken following the Blizzard of 1979 (the features are easier to see in the full size image listed above). This very strong winter storm even had a hurricane-like eye and dumped several feet of snow on the Midwest and along the East Coast. The snow cover on the ground is easily recognized, since it follows many geographical features. In the Appalachian Mountains of Pennsylvania and West Virginia, the snow follows the ridge and valley pattern. Along the Atlantic coast, many rivers cut through the snow-covered land. Other river valleys can be seen in the Midwest. Notice that snow-covered land in Canada shows up as a darker tone. This is because extensive coniferous forests block the snow from the satellite sensor. Farther north, in the tundra, the snow-covered land shows up as white, since there are no trees to block the snow.